Analysis of the Amino Acid Sequence Variation of the 67–72p Protein and the Structural Pili Proteins of Corynebacterium diphtheriae for their Suitability as Potential Vaccine Antigens

Corynebacterium diphtheriae is the etiological agent of a serious infectious disease – diphtheria. The diphtheria vaccine is highly effective but is directed only against the diphtheria toxin. The non-toxigenic C. diphtheriae strains may cause many severe invasive diseases, e.g., endocarditis, septic arthritis, bacteremia, and noninvasive wound infections. The notified increasing number of non-toxigenic C. diphtheriae infections indicates that infections are a growing problem in Europe (Belko et al. 2000; Zasada et al. 2010; Zasada 2013; Fricchione et al. 2014; Dangel et al. 2018). It is hypothesized that high vaccination coverage has resulted in the emergence of non-toxin-producing (non-toxigenic) C. diphtheriae strains, which acquired new virulence factors.

The diphtheria toxoid vaccination protects against the action of the toxin but does not protect against colonization and invasion by C. diphtheriae. Little is known about C. diphtheriae virulence factors other than the diphtheria toxin. The degree of adhesion of microorganisms to host cells has been shown to be an important pathogenicity factor for both toxigenic and non-toxigenic strains (Colombo et al. 2001). The virulence factors that facilitate bacterial colonization to specific host tissues and are related to pathogenesis include pili and fimbriae (Reardon-Robinson and Ton-That 2014). These structures are covalently attached to the bacterial cell wall and are recognized by the related host receptors (Sauer et al. 2000; Rogers et al. 2011). The surface structures are potential candidates for the development of new vaccines and antimicrobial therapies, due to their significant role in pathogenesis (Maione et al. 2005; Soriani and Telford 2010). Pili are found in both Gram-positive and Gram-negative bacteria, albeit with different folding mechanisms (Thanassi et al. 1998; Ton-That and Schneewind 2004). The function of pili is not only the involvement in adhesion, but they also act as bacteriophage receptors and participate in DNA transfer, biofilm formation, cell aggregation, host cell penetration, and motility (Proft and Baker 2009).

There are three types of pili in C. diphtheriae strains (SpaA, SapD, and SpaH), each containing the LPXTG motif (Ton-That and Schneewind 2003; Gaspar and Ton-That 2006; Swierczynski and Ton-That 2006). The genes involved in the production of pili encode nine pili proteins, defined in the successive letters from SpaA to SpaI, and five sortases defined from SrtA to SrtE, which are organized in three separate clusters. The sixth sortase SrtF (class-D homolog) is now referred to as the housekeeping sortase, located in a different region of the chromosome (Ton-That and Schneewind 2003). C. diphtheriae, like other Gram-positive bacteria (e.g. Streptococcus pneumoniae, group A and B streptococci or Actinomycetes), have the gene encoding cysteine transpeptidase (sortase) conserved in the genome, which is necessary for the assembly of pili (Ton-That and Schneewind 2004; Marraffini et al. 2006).

Pili are composed of three proteins: the main subunit forming the stem of pili, and two smaller subunits located at the base and at the end of the pili, e.g., SpaA-type pili is structured in such way that the SpaA pili protein creates a stem, SpaC is located at the end of pili, while SpaB is located along the stem and at the base (Ton-That and Schneewind 2003; Mandlik et al. 2008; Rogers et al. 2011). SpaA is important for the formation of the pile structure (Ton-That and Schneewind 2004). It has been proven that in the absence of SpaA protein, SpaB and SpaC are anchored in the cell wall as monomers (Mandlik et al. 2007).

The adhesion process of C. diphtheriae strains to the surface of human cells is multifactorial. Functions and mechanisms of action of fimbriae (Mandlik et al. 2007), non-fimbrial 67–72p adhesin (Colombo et al. 2001), trans-sialidase (Mattos-Guaraldi et al. 1998), hydrophobins, and sugar residues (Mattos-Guaraldi et al. 1999a; 2000; Moreira et al. 2003) are poorly understood, especially how they jointly participate in the adherence to the host cells and in the colonization of these cells during bacterial infection.

Initially, the 67–72p adhesive protein was described as a ligand responsible for the adherence of C. diphtheriae to human erythrocytes (Colombo et al. 2001). Later, the participation of this protein in the adherence of bacteria to HEp-2 cells was described (Hirata et al. 2004). The presence of the 67–72p protein has been confirmed in C. diphtheriae strains isolated from various sources, e.g., on the surface of the cells of invasive HC01 strain isolated from the blood of a patient with endocarditis (Sabbadini et al. 2012).

C. diphtheriae strains exhibit cell surface hydrophobicity and autoaggregation. Thanks to these features, microorganisms avoid immune defenses and are able to survive on the surface of the skin and mucosal membranes (Mattos-Guaraldi et al. 1999b). In addition, it has been proven that the 67–72p protein has the ability to induce host cell death, giving a signal for apoptosis in the early stages of infection. The occurrence of 67–72p hemagglutinin is one of the characteristics of the potentially invasive strains because it can contribute to the cytotoxicity and apoptosis of the infected cells (Sabbadini et al. 2012).

In our study, we analyzed the nucleotide and amino acid sequences of the genes encoding pili proteins which contribute to bacterial adherence to host cells, and also the gene encoding 67–72p protein involved in adhesion, colonization, and induction of the cell apoptosis in the early stage of infection, which should be used in a preliminary research for the finding of new vaccine antigens.

The isolates from invasive diseases and wound infections were included in the study because a wound can be a portal of entry for invasive infections. Moreover, future vaccines against non-toxigenic C. diphtheriae should protect against all kinds of infections. The analyses conducted in this study showed the variability of C. diphtheriae strains in terms of the nucleotide sequence of the genes encoding the 67–72p protein (99.37–100% average similarity) and structural proteins of pili SpaC (38.94–99.97% average similarity), SpaI (81.39–100% average similarity), and SapD (63.78–100% average similarity). The 27/E strain, the only representative of the mitis biotype, was the most different from the other strains tested. This strain did not have the gene coding for the SpaI protein, while the average similarities of the nucleotide sequence of genes encoding the 67–72p, SapD, and SpaC proteins were 99.37%, 98.64%, and 38.94%, respectively when compared to the reference strain. Despite the significant variation, we managed to locate two fragments of the 67–72p protein (fragments No. 3 and 5), where the nucleotide sequences were 100% identical for all C. diphtheriae strains tested (the most conserved in the genome). In contrast, for C. diphtheriae biotype gravis we found as many as nine fragments located in the sequence of the gene encoding 67–72p protein (fragment No. 1, 3, 5) and the genes coding for the pili SapD (fragment No. 1), SpaI (fragment No. 2), and SpaC (fragment No. 2, 3, 7, 8), which were 100% identical in all nine strains tested (Table III).

The nucleotide sequences of the genes investigated were translated into the amino acid sequences. We revealed that the identified mutations resulted in a reading frame shift or were synonymous and nonsynonymous substitutions (Table IV).

In addition, we observed that a fragment of 1380 bp was inserted in the place of the gene encoding the SpaC protein in the 89/E strain (Table III). After sequencing the fragment No. 4 of spaC, we obtained the exact nucleotide sequence, which we compared to the available sequences in the GeneBank using the Nucleotide Basic Local Alignment Search Tool (BLASTn). In this way, we proved that the 1380 bp insertion sequence (IS) was transposed, which interrupted the continuity of the tested gene between 3072 bp and 3073 bp of the reference sequence of the NCTC 13129 strain. The result from the BLAST search revealed that the insertion fragment we detected was similar to the sequence of IS3 family transposase also identified in other C. diphtheriae strains but in different locations.

Analysis of the 67–72p, SpaC, SpaI and SapD protein affinity to MHC classes I and II, and linear B-cell epitopes in the first stage relied on the determination of the position of the proteins tested in the cell membrane and confirmation that all selected proteins were at least partially membranous or extracellular (Table V). Then, using the IEDB platform, it was observed that all proteins have high-affinity areas for MHC receptors of both classes and the fragments, which can be recognized by antibodies. The output for the prediction of the high- affinity MHC binding peptides is typically given either in the units of a predicted affinity (IC₅₀ nanomolar) or as a percentile score reflecting the relative affinity of a selected peptide compared with a universe of random sequences. According to Paul et al. (2013), there are four categories of percentile ranks: 1) 0–0.30; 2) 0.30–1.25; 3) 1.25–5.0; and 4) 5.0–15.0. Their study proved that four pools of predicted peptides derived from the first two categories (0–0.30; 0.30–1.25) were immunogenic but finally, the transgenic mice in their study recognized only one peptide pool from the first category (0–0.30) (Paul et al. 2013). We can say that the smaller the percentile rank value, the higher the affinity. As for the IC₅₀ value, according to the IEDB Solutions Centre guidelines, IC₅₀ < 50 designs very high affinity, IC₅₀ < 500 – high affinity, and IC₅₀ < 5000 means low affinity (Fleri 2013). Accepting even the threshold of cutting off the percentile rank below 1 or IC₅₀ below 50, we still could have at least 100 to several hundred regions with high affinity for each of the proteins (Table VI, Table VII).

Table VIII contains data on the number of alleles for which epitope that was found in the proteins analyzed. Overall, 40 HLA alleles of human origin encoded by HLA-A and HLA-B were selected for ProPred1. The HLA 7 from mouse s-derived (MHC-Db, MHC-Db revised, MHC-Dd, MHC-Kb, MHC-Kd, MHC-Kk, and MHC-Ld) was omitted. The Proteasome Filter and Immunoproteasome filters were included in the analysis and for both, the threshold score of 4% was used. The ProPred1 cut-off threshold was also set at 4%. At the ProPred, 51 alleles related to MHC class II were considered. These were HLA-DR alleles. These molecules were encoded by DRB1 and DRB5 genes containing HLA DR1 (2 alleles), DR3 (7 alleles), DR4 (9 alleles), DR7 (2 alleles), DR8 (6 alleles), DR11 (9 alleles), DR13 (11 alleles), DR15 (3 alleles), and DR51 (2 alleles). The cut-off threshold was set at 3%.

The target protein sequences were scanned for B-cell epitopes using the Bepipred Linear Epitope Prediction 2.0. The selected B-cell linear epitopes of the proteins analyzed are shown in Table IX.

The results obtained with the VaxiJen server also confirmed the possibility of using the proteins as antigens in vaccines (Table X).

The huge success of vaccination against diphtheria almost enabled the elimination of the disease in Europe and other developed countries. However, in many countries with high vaccination coverage, i.e. France, Italy, Switzerland, Germany, and Canada an increase in non-toxigenic C. diphtheriae infections has been observed. For example, Poland is a country where the last case of diphtheria was recorded in the year 2000 and where the vaccination level of over 95% is achieved (Zasada et al. 2010). The first case of non-toxigenic infection with C. diphtheriae biotype gravis was recorded in Poland in 2004, where this bacterium induced sepsis and endocarditis in a patient (Zasada et al. 2005) and since then, practically every year, several cases of invasive C. diphtheriae infections have been diagnosed. In northern Germany, the number of non-toxigenic C. diphtheriae infections increased from five in 2013 to 23 in 2016, and 24 in only the first half of 2017 (Dangel et al. 2018). In England and Wales, a dramatic increase in infections was recorded since 1986, peaking at almost 300 cases in the year 2000 (Edwards et al. 2011). These examples indicate that the development of a new vaccine against non-toxigenic C. diphtheriae infection is of very important and necessary demand.

In vaccine development, the potential virulence factors exposed on the surface of a pathogen are considered as suitable antigens for an effective acellular vaccine. It has been shown that pili of Gram-positive bacteria play a direct role in the pathogenesis. For example, studies on S. pneumoniae have proved that those strains, which have a pili island, adhere better to lung epithelial cells than do the strains that lack this island. In the invasive disease model, the piliated strain is more virulent and has a competitive advantage over the pili-negative strain after the mixed intranasal infection. Infection with the piliated strain induces a stronger inflammatory response and a higher level of the tumor necrosis factor in the bloodstream of mice, which may be due to the higher adhesion of the piliated bacteria to the cells involved in the innate immune response and their detection by host cell pattern-recognition receptors (Barocchi et al. 2006). The pili proteins are used as antigens in vaccines, for example in some acellular pertussis vaccines (Mosley et al. 2016).

The adhesion of Corynebacterium to host cells was observed for the first time for C. renale pili, which caused agglutination of trypsinized sheep red blood cells (Honda and Yanagawa 1974). It was not until more than thirty years later that Mandlik and colleagues identified adhesins, which were involved in adherence to pharyngeal host cells – the minor pilins SpaB and SpaC of C. diphtheriae (Mandlik et al. 2007). Subsequent studies showed that wild type C. diphtheriae cells bind to human lung epithelial, laryngeal, and pharyngeal cells, whereas mutants that lacked SrtA (i.e. they did not polymerize the SpaA-type pili) showed more than a 90% lower ability to adhere to human pharyngeal cells. Moreover, mutants that lacked only the major pili subunit, SpaA, showed a 10% reduction in adherence to these cells. In contrast, mutants that lacked either of the minor pilin subunits, SpaB or SpaC, showed a 70–75% reduction in adhesion. In addition, the latex beads coated only with SpaB or SpaC were sufficient to adhere to the host pharyngeal cells, while the SpaA-coated beads did not bind. SpaB and SpaC are anchored in the cell wall as monomers independent of the pilus structures. It is likely that the long pili mediate the initial attachment, while the monomeric pilins on the surface of the bacteria participate in the formation of an adhesion zone allowing the delivery of toxins and other virulence factors and may even play a significant role in host cell signaling (Rogers et al. 2011).

In addition to pili, C. diphtheriae produces the 67–72p protein located on its surface, which is involved in colonization, induction of apoptosis, and epithelial cell necrosis that were once attributed exclusively to the action of the diphtheria toxin (Sabbadini et al. 2012). This finding was also confirmed by Cerdeño-Tárraga et al. (2003) who sequenced the genome of the British clinical isolate (strain NCTC13129 biotype gravis – used in our study as the reference strain) and proved that the recent acquisition of pathogenicity factors went beyond the toxin itself and included the fimbrial proteins and adhesins. The 67–72p can act as an invasive and apoptotic protein for C. diphtheriae strains. The ability to penetrate, survive and induce apoptosis in epithelial cells may explain the endurance and dissemination of C. diphtheriae (Sabbadini et al. 2012). Proteins, which act as adhesins were also detected among other bacteria, e.g., the occurrence of the extracellular Eap protein was confirmed to be involved in colonization of eukaryotic cells by S. aureus strains (Haggar et al. 2003).

Our research is based on reverse vaccination. This method relies on the sequencing of pathogen genomes and determination in silico the most likely protective antigens prior to conducting experiments to prove this. Originally, this method was used to identify antigens as probable candidate vaccines against serogroup B meningococci (Christensen et al. 2013). In another study, Droppa-Almeida et al. (2018) used several available bioinformatics tools to design the efficient immunodominant epitopes for the development of the peptide vaccine against C. pseudotuberculosis for sheep and goats. Thanks to this research, it was possible to highlight the importance of bioinformatics software in the design of vaccines, especially in the identification of appropriate vaccine candidates, immune-informatics analysis and design of peptide vaccine from multi-epitopes (Droppa-Almeida et al. 2018). Bioinformatics tools present a lot of advantages, such as speed and low cost, so we used them at particular stages of the research.

First, we selected the gene encoding the 67–72p protein and the three pili genes spaC, sapD and spaI as the genes most frequently detected in various C. diphtheriae isolates as it was reported in our previous study (Zasada et al. 2012). In the present study, we found two fragments of 594 bp and 215 bp in the nucleotide sequence of the gene coding for the 67–72p surface protein, which were identical in all analyzed strains of C. diphtheriae, and in total, nine identical in C. diphtheriae biotype gravis strains sequences of the genes encoding 67–72p protein and structural proteins pili SpaI, SpaC and SapD (Table III). The sequence stability of these fragments represents a first step toward being the potential vaccine candidates. The analysis of amino acid sequences of these fragments confirmed that the proteins tested are located in the membrane or cell wall and have a large extracellular part (Table V).

An effective vaccine should induce a protective and long-lasting immune response. Therefore, we carried out analyses of the affinity of the tested proteins to MHC classes I and II and linear B-cell epitopes. MHC class I presents antigens to CD8+ T-cells and MHC class II presents antigens to CD4+ T-cells. The antigens, which are recognized by CD4+ and/or CD8+ T-cell receptors, have the potential to stimulate a long-lasting and cytotoxic immune response. B-cell epitopes can induce both primary and secondary immunity. We showed that, in each of the proteins, areas with high affinity to MHC receptors can be distinguished (Table VI, Table VII, Table VIII) and we localized B-cell epitopes from target proteins (Table IX). In addition, the VaxiJen server was used that is a reliable and consistent tool for predicting protective antigens of bacterial, viral and cancer origin. The results obtained also confirmed that the proteins tested by us could be interesting to use as antigens in vaccines (Table X).

Our studies have shown that in the genome of the 89/E strain, the insertion element of 1380 bp was transposed and attached to fragment 4 of the gene encoding the SpaC pili protein. The process of transposition of ISs can inactivate genes (Trost et al. 2012). Mandlik et al. (2007) confirmed the reduced C. diphtheriae adhesive activity as the result of mutations at the base pili protein SpaB and at the tip pili protein SpaC of the SpaA-type pili. Premature stop codons in the continuity of the genes encoding the proteins responsible for the adhesion of bacteria to host cells inactivate them and limit the colonization process. We did not investigate the adhesive activity of the strain 89/E and, therefore, we can only posit the influence of the insertion on the adhesive properties of the strain based on the data published by other researchers.

Due to the fact that in many European countries the number of infections with non-toxigenic C. diphtheriae strains has recently increased, a key aspect of our research was the understanding of virulence factors other than the diphtheria toxin and identification of new vaccine targets. An important problem is also the understanding of the colonization process and in particular the mechanism of adhesion and structure of the proteins, which participate in this process. Blocking the 67–72p surface protein or pili structural proteins could effectively prevent the adhesion of C. diphtheriae bacteria to host tissues, colonization, and infection development. Due to the comparison of the nucleotide sequences of the C. diphtheriae strains identification of the most conserved sequences in the genome and determination of the variability between strains was achieved. The conserved sequences identified in 67–72p, SpaC, SapD and SpaI in our study are identical for all C. diphtheriae strains tested and contain the epitopes for B-cells and T-cells and will be used in further research on the construction of a new vaccine. The main limitation of this study is the small number of isolates investigated. However, the results obtained here support further studies with a larger number of isolates from different countries. Moreover, the results of in silico analysis should be confirmed by in vivo studies on an animal model. The new vaccine will act to inactivate the antigens responsible for the host cell colonization by C. diphtheriae strains and inhibit the development of infection.